# Smoothing the Disentangled Latent Style Space for Unsupervised Image-to-Image Translation

Source: [Smoothing the Disentangled Latent Style Space for Unsupervised Image-to-Image Translation](<https://www.sota2.com/research/paper/68fa6a261e209131cea265f4>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T23:17:37.192Z

Authors: Yahui Liu, Enver Sangineto, Yajing Chen, Linchao Bao, Haoxian Zhang, Nicu Sebe, Bruno Lepri, Wei Wang, Marco De Nadai

Published: 2021-06-16T17:58:21.000Z

## Abstract

Image-to-Image (I2I) multi-domain translation models are usually evaluated also using the quality of their semantic interpolation results. However, state-of-the-art models frequently show abrupt changes in the image appearance during interpolation, and usually perform poorly in interpolations across domains. In this paper, we propose a new training protocol based on three specific losses which help a translation network to learn a smooth and disentangled latent style space in which: 1) Both intra- and inter-domain interpolations correspond to gradual changes in the generated images and 2) The content of the source image is better preserved during the translation. Moreover, we propose a novel evaluation metric to properly measure the smoothness of latent style space of I2I translation models. The proposed method can be plugged into existing translation approaches, and our extensive experiments on different datasets show that it can significantly boost the quality of the generated images and the graduality of the interpolations.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Image-to-Image Translation on CelebA-HQ gender as class (test)](<https://www.sota2.com/research/sota/image-to-image-translation-on-celeba-hq-gender-as-class-test.md>): Image-to-Image Translation; CelebA-HQ gender as class (test); TC: 13.315.

- [Image-to-Image Translation on AFHQ 3 classes (test)](<https://www.sota2.com/research/sota/image-to-image-translation-on-afhq-3-classes-test.md>): Image-to-Image Translation; AFHQ 3 classes (test); TC: 3.462.

- [Image-to-Image Translation on AnimalFaces-10](<https://www.sota2.com/research/sota/image-to-image-translation-on-animalfaces-10.md>): Image-to-Image Translation; AnimalFaces-10; mFID: 38.93.

- [Image-to-Image Translation on Food-10](<https://www.sota2.com/research/sota/image-to-image-translation-on-food-10.md>): Image-to-Image Translation; Food-10; mFID: 61.13.

- [Image-to-Image Translation on CelebA-HQ](<https://www.sota2.com/research/sota/image-to-image-translation-on-celeba-hq.md>): Image-to-Image Translation; CelebA-HQ; FID: 35.93.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2106.09016>)

[PDF](<https://arxiv.org/pdf/2106.09016.pdf>)